AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

Graph-GRPO: Training Graph Flow Models with Reinforcement Learning

Baoheng Zhu, Deyu Bo, Delvin Ce Zhang, Xiao Wang

ICML 2026 · Graph learning › Graph algorithms and clustering · approach: Continuous DAG structure learning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RL training (Graph-GRPO) to the DeFoG base model (ablation)DeFoG · PMO (23 tasks)AUC-top10: 11.079 -> 16.901Improved
add refinement without RL (ablation)DeFoG · PMO (23 tasks)AUC-top10: 11.079 -> 15.251Improved
add refinement on top of RL (ablation)Graph-GRPO · PMO (23 tasks)AUC-top10: 16.901 -> 17.950Improved
add dynamic prior on top of RL and refinement (ablation)Graph-GRPO · PMO (23 tasks)AUC-top10: 17.950 -> 18.987Improved
add prescreening on top of RL, refinement and dynamic prior (ablation)Graph-GRPO · PMO (23 tasks)AUC-top10: 18.987 -> 19.270Improved
add RL training to DeFoG on TreeDeFoG · TreeV.U.N. (%): 73.5 -> 97.5Improved
add RL training to DeFoG on PlanarDeFoG · PlanarV.U.N. (%): 95.0 -> 95.0No clear effect
add RL training to DeFoG on PlanarDeFoG · Planarratio to train set: 3.2 -> 1.5Improved
use refinement instead of de novo generation with RL (ablation)Graph-GRPO · Valsartan SMARTSaverage top-10 scoreImproved
use refinement instead of de novo generation with RL (ablation)Graph-GRPO · Scaffold Hoppingaverage top-10 scoreNo clear effect
use Graph-GRPO instead of GDPO on braf docking scoreGraph-GRPO · Protein docking braftop-5% docking score: -11.197 -> -11.141Hurt
use Graph-GRPO instead of GDPO on parp1 hit ratioGraph-GRPO · Protein docking parp1hit ratio (%): 9.814 -> 60.763Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy